EXTRACTING CORRELATION CURVES FOR LIGHT EMITTING DEVICE

EXTRACTION DE COURBES DE CORRELATION POUR DES DISPOSITIFS LUMINESCENTS

English Abstract

Disclosed is a technique to extract the dynamic correlation in electrical and
optical behaviour that is
representative of display aging .

Description

CA 02692097 2010-02-04
FIELD OF THE INVENTION
The present invention generally relates to improving the spatial and/or
temporal non-uniformity of a
display.
SUMMARY OF INVENTION
The disclosed techniques provide accurate measurement and retrieval of the
correlation in electrical
and optical characteristics of the display as it ages and ways of applying
this information to improve
display uniformity.
ADVANTAGES
It helps improve the display uniformity and lifetime despite instability and
non-uniformity of individual
devices and pixels.

CA 02692097 2010-02-04
This technique can be applied to any type of display, including active matrix
organic light emitting diode
(AMOLED) displays.
Here, device samples (e.g. OLEDs) are located on different parts of display
active area as shown in Figure
1. Each sample is being stressed with a chosen stress condition that
represents different types of stress
that a pixel may undergo during the course of the display lifetime for that
application. One can stress
several OLEDs at the same stress levels and use a polling-averaging technique
to avoid defects,
measurement noise and other issues that may rise during the stress conditions.
The extracted data from the OLED samples can be used to update the predefined
correlation curves in
the system for compensation of, e.g. degradation in OLED efficiency.
Since the reference samples are limited and may not cover all stress
conditions, one can use a moving or
continuous averaging of stress for each pixel in the display and map out the
stress conditions of each
pixel as a linear combination of several reference OLEDs. This way, the
correlation curve can be tuned
for all pixels in the display to avoid errors associated with convergence in
the correlation curves due to
different stress conditions. For example, if there are two reference curves
for high and low stress (thigh
and flow), we can calculate a close curve for a pixel as following:
St(ti) _ (St(ti_1)*kavg+L(ti))/(kavg+1);
Khigh = St(ti)/Lhigh; Klow = {1- Khigh;
Kcomp = Khighfhigh(Al)+KIow//ow(Al);
Here, Kavg is the moving average coefficient, Khigh and Klow the respective
weight of fhigh and flow, Kcomp the
compensation factor, L(ti) the luminance of the current frame, Al is the OLED
current change for a fixed
voltage (note that this parameter can be replaced with a OLED voltage change
with a fixed current as
well). The Khigh and K10 can also be determined as a non-linear function of
the moving average of stress,
St(ti), as opposed to linear, if the non-linear map follows the aging models
better than linear.
The compensation factor, Kcomp can be used for compensation of the OLED
optical efficiency aging. Also,
one can change the Kavg based on the aging of the OLED. As the OLED ages more,
the divergence
between the two coloration curves increases. Thus, Kavg should be larger to
avoid a sharp transition
between the two curves. One can use Al to adjust the Kavg to improve the
performance. We call this
technique, dynamic moving averaging.
To further improve the extraction of the correlation curves for each OLED, one
can reset the system
after every aging step (or after user turns on or off the system). We call
this technique 'event-based
moving averaging'.
Kcomp=Kcomp_evt+ Khigh(fhigh(0l)fhigh(Dlevt)) ~Klow(flow(~l)flow(~levt));

CA 02692097 2010-02-04
Here, Kcomp_evt is the compensation factor calculated at previous event, and
Aleut is the change in OLED
current during the previous event (note that this parameter can be replaced
with a OLED voltage change
under fixed current as well).
Figure 2 shows a stress profile of a pixel for an arbitrary time period. Here,
the peak brightness is
assumed to be 500 nits, and we have the coloration curves for 500-nit and 30-
nit stress conditions.
Figure 3 shows the results of different calculation methods (moving averaging,
dynamic moving
averaging, and event-based moving averaging). Based on OLED behavior, one of
the above techniques
can be used to improve the compensation for OLED efficiency degradation.